Thomson Reuters, 169 Saxony Road, Encinitas, CA 92024, USA.
Expert Opin Drug Metab Toxicol. 2011 Mar;7(3):287-98. doi: 10.1517/17425255.2011.553191. Epub 2011 Jan 22.
Despite rapid progress in OMICs and computational technologies in compound safety assessment, drug failure rate due to toxicity is still unacceptably high. One reason for this is an inadequate interpretation of high-throughput preclinical data. Another reason is the poor mechanistic understanding of drug side effects as currently just a few compound targets are linked to specific adverse reactions.
Current performance issues with statistical analysis of OMICs data or gene/protein/compound lists are discussed, illustrating potential advantages of knowledge-based approaches in prediction of human toxicity. The authors show several examples of quantitative functional analysis, including cross-tissue toxicity predictions and integrated analysis of different types of OMICs data. They also describe novel approaches linking compound targets and associated pathways to side effects. The reader will gain an update on the recent developments in knowledge-based analysis in toxicogenomics and computational methods correlating protein targets with adverse reactions.
Quantitative pathway analysis is a useful approach for deriving multi-variant predictive biomarkers for drug safety. However, more comprehensive studies are needed for direct comparison of performance between pathway- and gene-centric methods.
尽管在组学和计算技术在化合物安全性评估方面取得了快速进展，但由于毒性而导致药物失败的比率仍然高得令人无法接受。原因之一是对高通量临床前数据的解释不够充分。另一个原因是对药物副作用的机制理解不足，因为目前只有少数几个化合物靶点与特定的不良反应有关。
讨论了当前组学数据或基因/蛋白质/化合物列表的统计分析中的性能问题，说明了基于知识的方法在预测人类毒性方面的潜在优势。作者展示了几种定量功能分析的例子，包括跨组织毒性预测和不同类型的组学数据的综合分析。他们还描述了将化合物靶点和相关途径与副作用联系起来的新方法。读者将了解毒代基因组学中基于知识的分析和将蛋白质靶标与不良反应相关联的计算方法的最新进展。
定量途径分析是一种用于衍生药物安全性的多变量预测生物标志物的有用方法。然而，需要进行更全面的研究，以便直接比较途径和基因中心方法的性能。
